Recently, ultrafine particles (diameter .about.100 .ANG.) have been vigorously studied and developed from both a purely academic aspect for experimentally solving Kubo effect and the like and an industrial aspect for utilizing them as ornaments, magnetic memory elements and the like [Hayashi; J. Vac. Sci. Technol. A5(4), 1375 (1987)]. Materials thereof, however, are limited to simple metal such as gold, iron and the like or stable oxides such as alumina and the like, and thus there is no example indicating that ultrafine particles of compound semiconductors are produced. While metallic ultrafine particles are obtained by thermally evaporating metal in the atmosphere of an inert gas such as an argon gas and the like (pressure .about.1 Torr) wherein a mean free path is restricted, the inert gas only restricts the mean free path of metallic atoms and is not taken into the ultrafine particles in this case.
While most attempts to produce ultrafine particles (quantum boxes) of the compound semiconductors have encountered severe difficulties and the prospects for realizing such particles are very dim, there are some prospective techniques being investigated. One of them is a method to combine an ultra thin film epitaxial growth method such as Molecular Beam Epitaxy (MBE), metalorganic Vapor Phase Epitaxy (MOVPE) and the like with a local processing method (etching, doping, disordering) effected by Focused Ion Beam (FIB). There is also a potential technique wherein a selective wet etching is used in place of the FIB processing. In any case, while this method is to process one dimensionally quantized structure layered by the MBE to expand a quantized dimension, achieving processing to an accuracy of .about.100 A is yet difficult and greatly dependent on a development to be achieved hereinafter. Another method locally deposits compound semiconductors directly or indirectly on a suitable substrate using thin needle electrodes. This method which seems flexible and attractive at first glance has a problem in reliability and reproducibility because a substrate and a temperature causing a successful epitaxial growth must be selected in practice and further the thin electrodes must be operated under severe epitaxy conditions.
With respect to a handling method of ultrafine particules, a prior art manipulator is in its primitive stage. There is of course no embodiment for handling ultrafine particles of compound semiconductors and metallic ultrafine particles are only handled by such methods that fine particles evaporated in the above inert gas and deposited on a wall are gathered by a brush or the fine particles are transported onto a sample stage of an electron microscope through a micro-jet stream.
Zealous developments are in progress to improve the performance and degree of integration of a semiconductor laser. At present, however, a quantum well laser making use of a one dimensionally quantized and layered thin film structure cannot provide excellent threshold current density, line width and temperature characteristics because quantization is not effected in a direction parallel to the thin film and defects such as steps exist. There are some methods for reducing interface defects such as by carefully controlling conditions for a thin film growth effected by the MBE or the like. If an ideal interface is realized, the characteristics of laser will be improved as necessary. The fact, however, that the quantization is not affected in the direction parallel to the thin film as a principle must remain as a factor for reducing the laser characteristics.
With respect to the integration of a semiconductor laser, a general structure including a Fabry-Perot type resonator is not basically suitable. There is an example wherein a mirror face not inferior to a cleaved facet is created on a substrate by reactive ion beam etching and a Fabry-Perot resonator is fabricated using it. Nevertheless, it is apparent that a stepped structure caused by this processing is a large obstacle to the integration of other elements. A prototype laser [Distributed Feedback (DFB) laser] provided with an embedded diffraction grating in place of a reflection edge is fabricated with advantageous characteristics. A usual DFB laser, however, only feeds back light of a particular wavelength diffused from a diffraction grating and uses it to control an oscillation mode, and thus gain is not modulated.
Since three dimensional quantum boxes have a very sharp discrete energy level, the application thereof to the active layer of a semiconductor laser can provide high performance laser. To obtain the three dimensional quantum boxes, however, ultrafine particles with a diameter of hundreds of Angstroms or less must be prepared. It is very difficult to produce this size of ultrafine particles of compound semiconductors of Groups III-V (or Groups II-VI), and then none of the tests making use of the FIB processing has been successful. Further, even if these ultrafine particles are produced, a method for sizing and handling them based on a new principle and a manipulator are required because their size is too small.
A compound semiconductor laser is integrated (opto-electronic integrated circuit: OEIC) to be used in communication, data processing, opto-computer applications and the like. While an example of a prototype OEIC is reported, the degree of integration thereof is still very low and said to be retarded more than ten years than a Si technology. To increase the degree of integration of the OEIC, the semiconductor laser must be arranged to a structure without a Fabry-Perot type resonator. A DFB laser aiming at single mode oscillation, narrow line width and improved temperature characteristics in addition to the above arrangement is being developed. While the DFB laser does not require an edge mirror, it cannot yet provide micro laser and improved laser characteristics at the present stage.